The molar ratio of retinol-binding protein to transthyretin (RBP:TTR) has been proposed as an indirect method to assess vitamin A status in children with inflammation. Neither reference values nor appropriate cut-off point are available for adults. RBP, TTR and retinol were determined in plasma from 100 healthy adults and 31 low-risk surgical patients with no inflammatory response. RBP:TTR percentile distribution from 99 healthy adults with plasma retinol > or = 0.7 micromol/l was: 2.5th = 0.24; 5th = 0.31; 10th = 0.32; 25th = 0.41; 50th = 0.47; 75th = 0.54; 90th = 0.67; 95th = 0.78 and 97.5th = 0.81. In order to define a cut-off point, receiver operating characteristic (ROC) curve was constructed, using plasma retinol as gold standard. ROC curve was based on data from the 131 studied subjects, 11 of whom (8.4%) were classified as deficient on the basis of plasma retinol < 0.7 pmol/l. According to ROC curve criteria, RBP:TTR ratio was considered a good test, the area under the curve being 0.822, p < 0.001. A cut-off-point of < or = 0.37 is proposed to detect vitamin A deficiency in adults, since it allows reaching high sensitivity (81.8%), specificity (79.2%) and predictive value (79.4%). The proposed cut-off point falls between 13th and 14th percentiles.